1. Test ID: AER-DYN-006
2. a, Solution Submitted by: A. Hämäläinen, VTT Processes,
P.O.Box 1604, FIN-02044 VTT, Finland
Email: Anitta.Hamalainen@vtt.fi
Date: 3.8.2004
b, Reviewed by:
Date:
c, Accepted by:
Date:
3. Code or Program Applied: HEXTRAN - SMABRE
4. Short description of the Code:
The coupled code HEXTRAN-SMABRE is used for 3D core modelling of the hexagonal fuel lattice of VVER’s and 1D flow modelling of the cooling circuit and its heat structures.
SMABRE calculates the whole thermal-hydraulics of the loops and the core in a sparse geometry with non-iterative solution method, while the HEXTRAN core model solves neutronics, core flow and heat transfer iteratively. But additionally, HEXTRAN performs the detailed thermal hydraulics and fuel heat transfer calculation in every fuel assembly of the core to get the nodal fuel and coolant conditions for the calculation of three-dimensional neutron kinetics and reactivity feedback effects needed in it.
5. Known Approximations:
Perhaps the biggest restriction in the system codes generally, dealing with several kind of transients, is the numerical diffusion. Due to this, the accuracy of modelling in sharp changes, such as boron or temperature fronts is poor.
Modelling extensions are needed for a more accurate description of fuel assemblies. The heat transfer models used for the fuel assemblies in this benchmark don’t produce the burnup dependent pellet radial heat generation profile. In the benchamark the heat transfer coefficient in the gas gap was given and no better models were needed.
The four-equation model of HEXTRAN hydraulics pre-supposes that the steam phase is always saturated. Another restriction is that the used four-equation model cannot calculate the drying out of the liquid phase in the flow channels. The five-equation model of SMABRE has only a single momentum equation for the mixture. In both codes thermalhydraulics is solved one-dimensionally.
1 of 6

Spacers are treated as local pressure losses, with a constant loss coefficient. Local loss or gain due to abrupt area changes (e.g. at the upper end of the plenum of part length fuel) is modelled with a constant loss/gain coefficient given in input. All local losses are calculated with homogeneous two-phase multiplier.
6. Mathematical Model:
-
HEXTRAN
The three-dimensional core neutronics, heat transfer and thermal-hydraulics solution method of the HEXTRAN /1/ code is based on coupling and extension of the steady-state hexagonal core simulator HEXBU-3D /2/ and the one-dimensional thermal-hydraulics code TRAB /3/.
The two-group neutron diffusion equations are solved in HEXTRAN by a nodal expansion method in x-y-z geometry within the reactor core. A basic feature of the method is decoupling of the two-group equations into separate equations for two spatial modes and reconstruction of group fluxes from characteristic solutions to these equations. The two solutions are called the fundamental or asymptotic mode, which have a fairly smooth behaviour within a homogenized node, and the transient mode, which deviates significantly from zero only near material discontinuities.
The nodal equations are solved with a two-level iteration scheme where only one unknown per node, the average of fundamental mode, is determined in inner iterations. The nodal flux shapes are improved in outer iterations by recalculation of the coupling coefficients. Cross sections are computed from polynomial fittings to fuel and coolant temperature, coolant density and soluble boron density.
The cross sections for VVER applications were earlier calculated with the cell-burnup code
CASMO-HEX /4,5/ and recently with the hexagonal option of CASMO-4 /6/. Nodal distributions for fuel burnup and Xenon are obtained from the fuel management code
HEXBU-3D.
The thermal-hydraulic calculation of the reactor core is performed in parallel one-dimensional hydraulic channels, each channel usually coupled with one fuel assembly. Channel hydraulics is based on conservation equations for steam and water mass, total enthalpy and total momentum, and on a selection of optional correlations describing e.g. non-equilibrium evaporation and condensation, slip, and one and two-phase friction. The phase velocities are related by an algebraic slip ratio or by the drift flux formalism. The thermal hydraulic solution methods are the same as in the one-dimensional code TRAB /3,7,8/.
During the hydraulics iterations one-dimensional heat transfer calculation is made for an average fuel rod of each assembly. The radial heat conduction of the fuel rod is solved according to Fourier's law. Thermal properties of fuel pellet, gas gap and fuel cladding are functions of local temperature and burnup, and the heat transfer coefficient from cladding to coolant depends on the hydraulic regime. The fission power is divided into prompt and delayed power parts, and a fraction of the power can be dissipated into heat directly in the coolant. Decay heat is included in the thermal power.
2 of 6

Advanced time integration methods are applied in the dynamic calculation. The numerical technique can vary between the standard fully implicit theta method and the central-difference theta method both in the heat conduction calculation for fuel rods and in the solution of thermal-hydraulic conservation equations for cooling channels.
The core geometry consists of separated parallel one-dimensional flow channels. The channels can be further divided into axial sub-regions, to take into account the geometric, hydraulic and heat transfer characteristics of the fuel bundles. Parallel to the heated channels, several unheated by-pass channels can be modelled. One or several hot channels, with possible hot rods, can be modelled separately from the calculation of the whole core.
HEXTRAN is a best estimate code. Possibilities to modify the neutronics parameters have been included in the code so that the conservativity of the calculations can be simply and reliably modified without changing the ordinary neutronics data.
Some of the key non-proprietary hydraulics correlations available in HEXTRAN are:
 Baroczy’s correlation as applied by Chisholm for the two-phase multiplier
 Epri’s correlation for drift flux/slip calculation
 The nonequilibrium model for boiling/evaporation.
 McAdams correlation for distributed friction
-
SMABRE
The development of SMABRE /9,10/ code started in the beginning of 1980’ies. The development of the model originated from practical need for a fast running thermal hydraulic model for scoping studies of small break LOCA accidents. SMABRE model is based on a non-iterative algorithm of five conservation equations with a single momentum equation for the mixture. The phase separation is solved by using the drift flux model.
The selection of constitutive models is presented in Table 1. For the heat transfer simplified and modified correlations have been used, but for typical light water reactor cases the differences into the original correlations are quite small. The interfacial heat transfer coefficients are smaller than in many analysis codes and the reason is that the SMABRE model was derived as non-iterative solution, which should tolerate large time steps. The condensation correlation gives a reasonable condensation rate and the condensation is not strongly flow rate dependent. The flashing model calculates well enough the behaviour during blowdown conditions, but the flashing of the fast depressurisation is not considered.
Table 1. Constitutive thermal hydraulic models for SMABRE
Physical phenomena
Wall friction
Interfacial friction
Net vaporization
SMABRE model
Blasius equation for mixture
Not modelled for the drift flux model
A linearized ramp function from subcooled liquid to saturation point
Pre-DNB heat transfer Dittus-Boelter, Chen as simplified for boiling
Critical heat flux for wall Zuber-Griffith, VVER: Smolin, Bezrukow heat transfer
Post-DNB wall heat Dittus-Boelter to gas
3 of 6

transfer
Interfacial condensation Droplet type condensation or through stratified water level
Interfacial flashing Linear function of liquid mass and liquid superheat
Critical flow limitation Sound velocity limitation or Moody model applied for the junction
Pump characteristics Four quadrant curves for head and torque for flow and pump speed.
Phase separation
Material property solution
Drift flux model derived from EPRI correlation or full separation
Rational function fittings, two- or one parameter functions
The point kinetic model simulates one energy group for neutrons and 6 groups for delayed neutrons. Reactivity may be defined by simply reactor feedback coefficients or in the table form as in RELAP5. Reactivity feedback is calculated as a function of average liquid density and temperature, fuel average temperature and boric acid concentration in the core and the initial state feedback parameters define the equilibrium reactivity level.
The numerical solution for the SMABRE model is a predictor-corrector type non-iterative solution. The sparse matrix inversion is used for the solving of pressure, void fraction and energy distributions. The pressure solution includes implicitly the result of the flow distribution.
- Coupling of codes
The codes have been connected using parallel coupling. The coupled code HEXTRAN-
SMABRE has its own main program and a few interfacing subprograms, but in the combination HEXTRAN and SMABRE are used as if they were separate codes. Both use their own input, output, restart and plotting. HEXTRAN dictates the time step. SMABRE calculates the whole thermal-hydraulics of the loops and the core in a sparse geometry. But additionally, HEXTRAN performs the detailed thermal hydraulics and fuel heat transfer calculation in every fuel assembly of the core to get the nodal fuel and coolant conditions for the calculation of three-dimensional neutron kinetics and reactivity feedback effects.
The SMABRE core model typically consists of as many parallel sectors as there are loops in the plant, each divided into 5 to 10 axial nodes. In the HEXTRAN core model, each fuel assembly is normally divided into 20 to 25 axial nodes for thermal hydraulics, neutronics and heat transfer. Typically each assembly is associated with a separate flow channel, but several assemblies can also be combined to a flow channel.
In the combined code the interchanged variables between modules are the nodal power to coolant distribution from HEXTRAN to SMABRE, and core outlet pressure, inlet pressure, inlet mass flow and enthalpy and inlet boron concentration for each core sector from
SMABRE to HEXTRAN. The data is exchanged once during a time step.
7. Features of Techniques Used:
-
4 of 6

8. Computer, Operational System:
Tru64Unix HPUX
9. References:
1.
Kyrki-Rajamäki, R. Three-dimensional reactor dynamics code for VVER type nuclear reactors. Espoo: Technical Research Centre of Finland, 1995. 51 p. + app. 80 p. VTT
Publications 246. Dr.Tech. Thesis.
2.
Kaloinen, E., Teräsvirta, R. & Siltanen, P., HEXBU-3D, a three-dimensional PWR simulator program for hexagonal fuel assemblies. Espoo: Techincal Research Centre of
Finland, Nuclear Engineering Laboratory, 1981. 148 p. + app. 5 p. (Research Report 7).
3.
Rajamäki, M. TRAB, a transient analysis program for BWR, Part 1, Principles. Espoo:
Technical Research Centre of Finland, 1980. 101 pp. + app. (Nuclear Engineering
Laboratory, Research Report 45.) ISBN 951-38-0916-1.
4.
Wasastjerna, F., Cellutbränningsprogrammet CASMO-HEX / Anpassning av det svenska programmet CASMO till hexagonal geometri och vissa undersökningar beträffande dess noggrannhet. Licentiatarbete, Tekniska högskolan (Helsinki University of Technology), avdelningen för teknisk fysik, 1977.
5.
Anttila, M., The fuel assembly program CASMO-HEX. Input manual. Technical report
RFD-17/96. 5p.
6.
Rhodes, J., Smith, K., Edenius, M., CASMO-4E, Extended Capability CASMO-4, User’s
Manual. Studsvik Scandpower , proprietary SSP-01/401, 2001.
7.
Rajamäki, M., TRAWA, a transient analysis code for water reactors. Espoo 1976.
Technical Research Centre of Finland, Nuclear Engineering Laboratory, Report 24. 149 p
+ app. 31 p.
8.
Räty, H., Fuel heat conductivity and heat capacity dependencies in TRAB version 3.5 and
HEXTRAN version 2.7. Espoo 1996. VTT Energy, Nuclear Energy, Technical report
RFD-3/96. 5 p.
9.
Miettinen, J., 1985, Development and Assessment of the SBLOCA Code SMABRE,
Proceedings of CSNI Specialists' Meeting on Small Break LOCA Analyses in LWRs,
Vol. 2, p. 481, Università di Pisa, Pisa, Italy.
10.
Miettinen, J. & Hämäläinen, A. Development and Validation of the Fast Running
Thermohydraulic Model SMABRE for Simulator Purposes. In: Proceedings of ICONE-8:
Eight International Conference on Nuclear Engineering. Baltimore, USA, 2-6 April, 2000
[CD-ROM]. New York: American Society of Mechanical Engineers. Paper ICONE8-
8188, 12 p. ISBN 0791819922.
5 of 6

11.
Hämäläinen, Anitta & Kyrki-Rajamäki, Riitta, HEXTRAN-SMABRE calculation of the
6 th AER benchmark, main steam line break in a VVER440 NPP. Proceedings of the 13th symposium of the AER on VVER reactor physics and safety, September 22-26, 2003 in
Dresden, Germany. KFKI/AEKI, Budapest, 2003. Pp. 445 - 458. ISBN 963-372-630-1.
12.
Kliem, S., Comparison of the updated results of the 6th Dynamic AER Benchmark – main steam line break in a NPP with VVER440. Proceedings of the 13th symposium of the AER on VVER reactor physics and safety, September 22-26, 2003 in Dresden,
Germany. KFKI/AEKI, Budapest, 2003. Pp. 413 - 444. ISBN 963-372-630-1.
10. Results:
All requested results are in the ASCII-file dyn006_solvtt.txt
. The solution of the benchmark is described in /11/. The comparison with other solutions is presented in the specification document dyn006.pdf
and in /12/.
11. Comparison to Recommended Solution:
No reference solution does exist. The comparison with solutions of other coupled codes is presented in the specification document dyn006.pdf
and in /12/.
6 of 6